Google's Gemini 3.8 Flash Works Harder, Costs More for Developers

Google has released Gemini 3.8 Flash, a new version of its AI model that 'works harder' and delivers more accurate results by running more reasoning steps on complicated prompts.

The catch is that this extra effort comes at a cost: developers may need to pay more tokens for the same job, even though the price per token hasn't changed since the last version.

Google has framed the shift in costs as a subtle but meaningful change in how it prices its 'Flash' tier models. The company acknowledges that developers who want predictable and minimal token usage may prefer to stick with Gemini 3.7 Flash, which remains available.

The release of Gemini 3.8 Flash marks a breakneck pace in the AI industry, where companies like OpenAI, Anthropic, and Google are pushing smaller, faster models that still punch above their weight on reasoning tasks.